--- Log opened Fri Jan 18 00:00:42 2013
00:03 < jjjjohnnny> is there a repo of browserifiable front end modules yet?
00:03 < mbalho> that would be a big repo
00:04 < mbalho> raynos has a gist or something
00:10 < substack> yay signups
00:10 < mbalho> mikolalysenko: there were people working on node kalman filters at a drone meetup i went to recently, not sure if they ever published it though
00:11 < mbalho> guybrush_: i dont do prediction yet, gonna hack on that soon
00:11 < mbalho> guybrush_: i just lerp server updates
00:11 < substack> jjjjohnnny: all these repos have testling tests which is a really good indicator that they work with browserify http://ci.testling.com/commits.json
00:12 < substack> the commits.json data is fed into this: http://browserify.org/search
00:12 < substack> the best way to solve this problem is to prove that modules work in browsers by writing tests demonstrating as much!
00:13 < guybrush_> mbalho: gotcha
00:14 < mbalho> substack: heres an interesting writeup, check out the 1.2.3.4 https://sites.google.com/site/letsmakeavoxelengine/home/collision-detection
00:16 < substack> mbalho: oh sweet
00:17 < substack> need a reusable module for this stuff
00:17 < substack> with independent tests
00:17 < mbalho> yea
00:18 < mbalho> substack: also http://stackoverflow.com/questions/4694759/collision-detection-by-sliding-against-a-plane-in-xna
00:21 < chrisdickinson> i think the tile based thing i'm working on might prove useful
00:22 < chrisdickinson> (basically, walk each axis separately in the direction of movement, walking the other two axises in a predetermined fashion)
00:22 < chrisdickinson> then if there's a collision on an axis, modify the vector on that axis
00:22 < chrisdickinson> then walk the other axises
00:23 < chrisdickinson> the biggest part is walking the other two access in a reliably "clockwise" fashion
00:23 < mbalho> i just want collisions to be solved
00:23 < mbalho> :D
00:24 < chrisdickinson> for real
00:24 < chrisdickinson> bonus points, if this collide-3d module works it'd be really easy to plop it into any number of other minecraft-ish simulations
00:24 < mbalho> yea
00:34 < substack> plop
00:35 < jjjjohnnny> neat
00:41 < chrisdickinson> oooh
00:42 < chrisdickinson> game.controls.velocity isn't in world space yet
00:42 < chrisdickinson> mbalho: is there a worldspace version of the velocity vector?
00:43 < chrisdickinson> or is there a way to get that vector?
00:44 < mbalho> chrisdickinson: hmm i dunno
00:44 < mbalho> chrisdickinson: you could pass it through Chunker.prototype.voxelVector maybe
00:45 < jez0990_> Hey folks. Where's a good place I can discuss philosophy with other programming geeks like us?
00:46 < mbalho> your local hackerspace
00:46 < jez0990_> you mean I have to leave my house?? :(
00:57 < chrisdickinson> wooo it works but really slows down
00:58 < mbalho> lol
00:59 < chrisdickinson> like it slides around boxes and everything
00:59 < chrisdickinson> but the timestep is wrong, somehow
00:59 < chrisdickinson> mbalho: i'll be online later and put up a demo
01:00 < mbalho> chrisdickinson: cool!!
01:02 < chrisdickinson> it's kind of ugly since i need to get a world vec then reapply it to the player
01:03 < ryanseddon> can anyone help me, have I done anything wrong to get testling-ci working https://github.com/ryanseddon/H5F/commit/975e8fbfe524e88160c53f2c912acc92b5adae0f
01:05 < substack> ryanseddon: we were getting a bug pushing your repo with github-push-receive :/
01:06 < ryanseddon> oh so i didn't do anything wrong or something i did made it fail?
01:07 < substack> it's a bug on our end we haven't figured out yet
01:07 < ryanseddon> substack: ok cool so everything in my commit looks good to you otherwise?
01:07 < substack> well not sure yet because of the bug
01:08 < ryanseddon> right, well i'
01:08 < ryanseddon> right, well i'll be lurking so let me know if I need to change anything
01:15 < jez0990_> mbalho: here's an idea for voxel.js - mmo (or p2p mmo) works by clients (or web-rtc'd peers) connecting their meta-voxel worlds through adjacency in two dimensions
01:15 < jez0990_> how much would you say is left to build before that is feasible?
01:21 < Raynos> mbalhbo: ?
01:22 < Raynos> jjjjohnny: https://gist.github.com/4350728
01:24 < Raynos> please mention any extra modules to be added in a comment
01:26 < jez0990_> Raynos: these lists are awesome by the way, not enough people publish them - thanks!
01:45 < jez0990_> chrisdickinson: how did the dates on your neversaw.us blog manage to get two years out of sync?!
02:00 < substack> Raynos: do you have a module for taking a readable/writable stream and only making it writable?
02:38 < Raynos> substack: no
02:38 < Raynos> substack: but (stream.read = null && stream.pipe = null) should work
02:38 < Raynos> substack, mbalho: whats this event tongiht at 8pm?
02:39 < substack> Raynos: http://www.5mof.net/
02:39 < Raynos> substack: actually `return WriteStream(stream.write, stream.end)` might work
02:41 < Raynos> substack `npm docs write-stream`
02:41 < mbalho> jez0990_: i dont know enough about webrtc to make it happen but i am using scuttlebutt for world data sync in the voxel-server demo
02:41 < mbalho> jez0990_: so if you can get scuttlebutt to work over webrtc it can happen
02:41 < Raynos> mbalho, jez0990_: https://github.com/Raynos/painter/blob/master/examples/paint/index.js#L47
02:42 < Raynos> that's not real webrtc though. That's emulated webrtc
02:42 < mbalho> yea i want real webrtc
02:45 < Raynos> mbalho: real webrtc is nightly builds of firefox / chrome and it's UDP and there's no firefox - chrome interop
02:45 < Raynos> The nice thing about the emulate webrtc is that it has the same API
02:45 < Raynos> once chrome beta channel has data channels ill make it work
02:46 < mbalho> ah
02:47 < Raynos> https://github.com/Raynos/peer-connection-shim <- that's a RTCPeerConnection object which uses relays
02:54 < guybrush_> substack, is glog broken currently? i get an error when i push on glog/index.js:53 TypeError: Object blog.git has no method 'once'
02:54 < guybrush_> though didnt read any code yet or looked deeper
02:54 < substack> it's running browserify.org
03:08 < substack> guybrush_: maybe the pushover version is out of sync?
03:08 < guybrush_> 0.1.6
03:10 < guybrush_> oh i think npm i pushover@latest fixed it
03:13 < chrisdickinson> jez0990_: it's statically generated, and i haven't paid any attention to it :)
03:17 < substack> jjjjohnnny: rec -t s16 -
03:19 < guybrush_> ah sweeeet yes glog is cool
03:19 < guybrush_> pretty much best blogging-thing ever
03:20 < guybrush_> though i think this has to be upgraded https://github.com/substack/glog/blob/master/package.json#L10
03:22 < substack> it should install the latest pushover on a fresh install
03:26 < chrisdickinson> mbalho: you around?
03:31 < chrisdickinson> mbalho: http://neversaw.us/voxel/ so, be careful around seams (wherever voxel chunks are joined), they're still wonky
03:31 < chrisdickinson> jump physics are a bit jerky
03:31 < chrisdickinson> but the voxel collisions seem pretty solid otherwise
03:32 < chrisdickinson> welp maybe not lol.
03:32 < chrisdickinson> still a few bugs :D
03:38 < substack> chrisdickinson: a bit fast
03:39 < chrisdickinson> yeah
03:39 < substack> chrisdickinson: whoa is that your collision detection?
03:39  * chrisdickinson nods
03:39 < chrisdickinson> it's definitely buggy
03:39 < substack> that is so good
03:39 < substack> soooo much better than my crappy thing
03:40 < chrisdickinson> it's basically https://github.com/chrisdickinson/collide-2d-tilemap in 3d
03:40 < substack> sweet
03:40 < chrisdickinson> also thanks!
03:40 < chrisdickinson> there are definitely some cases that are… very very broken
03:40 < chrisdickinson> seams, and some weirdness with being in tight spaces
03:40 < substack> feels like the field of view is too narrow
03:41 < substack> oh and the camera is on the floor
03:41 < substack> not at eye-level
03:41 < substack> the position should be up at 1.5 block units or so
03:41 < substack> for the player camera I mean
03:41 < substack> that's the biggest reason why it feels weird I think
03:41 < chrisdickinson> ah
03:42 < chrisdickinson> lemme check in firefox really quick
03:43 < substack> I'm getting teleported a strange distance away when I fall in a hole
03:43 < chrisdickinson> yeah
03:43 < chrisdickinson> that's the biggest issue
03:43 < chrisdickinson> well, aside from seams
03:43 < defunctzombie> substack: your resolve module, the basedir that I give it, what is that meant to be? do I need to have node_modules folder at that location?
03:43 < defunctzombie> or up from that location?
03:43 < chrisdickinson> i'll push up the changes to a fork and publish the modules
03:44 < chrisdickinson> (collide-3d-tilemap and aabb-3d)
03:44 < substack> defunctzombie: the basedir is just the directory from which you want the path resolution algorithm to start
03:44 < substack> because require.resolve() is always in the __dirname
03:44 < defunctzombie> gotcha, and it will go up from there right?
03:44 < substack> so you can't pass a dirname as an argument
03:44 < substack> defunctzombie: exactly
03:44 < defunctzombie> k, thx
03:45 < defunctzombie> cause I am trying to resolve something based on that notion
03:45 < defunctzombie> but it can't seem to find it
03:45 < defunctzombie> oh
03:45 < defunctzombie> fucking relative paths
03:51 < mbalho> im about to give a talk on voxeljs then ill look at collisions
03:52 < substack> mbalho: I started a portal gun repo and once I get this tap-parser thing working I'll hack on it a little more
03:55 < chrisdickinson> substack: how are you implementing physics on your end?
04:00 < defunctzombie> Raynos: https://github.com/shtylman/npm-css/blob/master/index.js
04:00 < defunctzombie> yes, I know I changed quite a bit of stuff, but I made tests actually start working :)
04:00 < substack> chrisdickinson: seeing if points along a rectangular prism are inside the geometry
04:00 < defunctzombie> and figured this could be some sort of starting point
04:00  * chrisdickinson nods
04:01 < substack> chrisdickinson: you should be able to just pos.y -= self.cubeSize * 1.5 to boost the player height up enough
04:01 < defunctzombie> Raynos: it is all sync currently (too many parts of this are sync right now to be async efficiently)
04:03 < Raynos> :p
04:03 < Raynos> you changed the code
04:03 < Raynos> from a complex chain
04:03 < Raynos> to a single function
04:03 < mbalho> substack: hellllll yeaaaaa portals!!
04:04 < fotoverite> awesome dominictarr
04:07 < chrisdickinson> https://github.com/chrisdickinson/aabb-3d
04:10 < defunctzombie> Raynos: yes
04:11 < defunctzombie> you are correct in that I made the code simpler :p
04:11 < chrisdickinson> substack: http://npm.im/aabb-3d and https://npmjs.org/package/collide-3d-tilemap
04:11 < defunctzombie> and actually work
04:11 < defunctzombie> :D
04:11 < substack> hooray!
04:12 < substack> chrisdickinson: planning on ripping out the collision detection in voxel-engine?
04:12 < chrisdickinson> ideally
04:13 < chrisdickinson> at least for colliding with the voxel map itself!
04:14 < chrisdickinson> object-object collisions could be handled by more traditional aabb-aabb collisions
04:21 < defunctzombie> Raynos: let me know what you think of the changes and such, ideally I am good with that package name and like the approach
05:22 < substack> Raynos, defunctzombie: tape@0.2.0 now has an `at` field https://gist.github.com/4562523#file-gistfile1-txt-L15
05:23 < defunctzombie> substack: is that where the assert was called from that failed?
05:23 < substack> yes
05:24 < defunctzombie> substack: does it also capture in the browser? or is this node side?
05:25 < substack> it should work in chrome and maybe ff
05:25 < substack> it just parses the err.stack from a new Error
05:25 < defunctzombie> ah
05:25 < substack> for the first file not in __dirname
05:25 < substack> it'll just skip setting this in browsers that don't have err.stack
05:26 < guybrush_> substack: so with glog every article is a tag with 1 file? or can i push multiple files and `glog publiish file` is just the entry-file?
05:26 < substack> guybrush_: glog publish just creates tags
05:26 < guybrush_> im pretty sleepy and lazy so i just ask instead of reading the code :P
05:26 < substack> it doesn't push
05:27 < guybrush_> ahhh i see
05:27 < guybrush_> so i just push everything, like folders pictures and stuff
05:27 < guybrush_> and than i glog publish which creates tags
05:28 < substack> sure
05:28 < guybrush_> niiice
05:28 < defunctzombie> substack: still not sure I buy into the whole always output tap stuff :/ kinda think the runner should just do whatever it wants and output tap if it wants
05:28 < defunctzombie> I thought about the notion of plan and such
05:28 < guybrush_> git-database for blogs is just awesome
05:28 < substack> guybrush_: I'm not sure if ecstatic serves the static content from the git repo yet though for pictures
05:28 < substack> you might need to hack that in and send me a pull request
05:28 < guybrush_> well i can hack arround that anyway
05:28 < defunctzombie> but not sure it is really that important :/
05:28 < guybrush_> sure
05:29 < substack> defunctzombie: sure but then TAP gets neglected
05:29 < defunctzombie> you can just have any assertion whatever throw in a test (and since tests are run in series) you know waht test just failed
05:29 < substack> and then tooling because crappier
05:29 < substack> that's what happened in mocha
05:29 < substack> I would rather just have only TAP so it gets first-class treatment
05:29 < substack> s/because/becomes/
05:30 < substack> defunctzombie: throwing is really bad for browser tests
05:30 < defunctzombie> is it? you can just do onerror right?
05:30 < substack> don't throw please
05:30 < substack> defunctzombie: that is pretty unreliable
05:30 < defunctzombie> ah, ok, fair enough. I was under the impression it was reliable
05:30 < substack> throwing too often breaks the transport layer to get the data out of browsers
05:31 < defunctzombie> transport layer?
05:31 < guybrush_> woah that sucks
05:31 < substack> so the browsers just go completely dark
05:31 < guybrush_> just shim assert right?
05:31 < substack> because throwing stops the javascript execution context
05:31 < defunctzombie> that is unfortunate indeed :(
05:31 < substack> you can do it in node because you get an exit code
05:31 < substack> but there is no such thing in browsers
05:31 < substack> just don't assert things and you'll be a lot better off
05:31 < defunctzombie> and no domains either to try and handle that issue :(
05:32 < substack> yes
05:32 < substack> another problem with not treating TAP as first-class is that your tests won't map to TAP output 1:1
05:33 < defunctzombie> so do some browsers just not continue script execution after anything throws?
05:33 < defunctzombie> that is ghetto, but understandable from their standpoint I guess
05:33 < defunctzombie> they just want to stop the script
05:33 < substack> so your tooling that consumes tap won't necessarily pick up on all the things that tap is designed to prevent
05:33 < substack> certain classes of false positives
05:33 < substack> defunctzombie: pretty much that's the case
05:33 < substack> it's really unpredictable what will happen
05:34 < substack> even if you have an onerror handler
05:34 < substack> it's really context-dependent on what the code exactly does
05:35 < defunctzombie> gotcha
05:35 < substack> cool, disabling css acceleration seems to have fixed some of my browsing woes ever since changing some chrome flags to get webgl working
05:42 < defunctzombie> substack: does tape do test timeouts or global var leak detection?
05:48 < substack> defunctzombie: no
05:49 < substack> out of scope I think
05:50 < defunctzombie> hard to say, I can see both sides of it
05:50 < defunctzombie> ideally static analysis can catch the common leaks
05:50 < defunctzombie> but that might not be true in some instances
05:51 < defunctzombie> timeouts I can also see both ways since the runner is in a good position to provide a nice interface for that
05:53 < defunctzombie> for me, testing is about ensuring my program is behaving properly
05:53 < substack> use a different tool for those
05:53 < defunctzombie> generally that means not leaky and runs within certain time constraints as side effects
05:53 < substack> a runner program like `tap` should care about timeouts
05:53 < substack> and another tool entirely should concern itself with detecting global leaks
05:54 < defunctzombie> tape does not act as a runner? just the harness?
05:54 < defunctzombie> it doesn't come with a tool for doing any sort of running?
05:54 < substack> tape only outputs TAP data to console.log()
05:54 < substack> that is its job
05:55 < substack> it does some rudimentary other kinds of things but I would like to limit the scope of those activities
05:56 < defunctzombie> how does that interface with tap then?
05:56 < defunctzombie> if tape is defining the test function
05:56 < defunctzombie> it is controlling how the test are run I presume
05:58 < substack> they interface via process.stdout
05:58 < substack> if you use `tap` to run your tests
05:59 < defunctzombie> so tap will just exit for long running tests and test timeout conditions (or a tool like tap) ?
06:00 < defunctzombie> if tape is doing the output, would I ever see test timings even
06:05 < substack> see test timings?
06:06 < substack> even if the runner did nothing you would see that your test suite hung at a certain point in its output
06:06 < substack> I would think that should be sufficient for most cases
06:13 < defunctzombie> yea, that is true, could still be nice to have test runtime info in tap output for tracking
06:14 < defunctzombie> some fancy places track runtime of tests
06:16 < defunctzombie> substack: any particular reason you don't provide the full stacktrace and just provide 'at' ?
06:22 < substack> the whole stack trace has test harness internals in it
06:57 < st_luke> testing leveldb with gigantic disjoint-sets to see how fast it is
09:01 < jesusabdullah> leveldb sounds fun
09:27 < mirkok> dominictarr: if you are interested in the results of my canonical JSON.stringify solutions
09:27 < mirkok> https://github.com/mirkok/canonical-json
09:27 < dominictarr> cool
09:29 < mirkok> oh and interestingly native JSON.stringify is only twice as fast as crockford's implementation in pure js
09:30 < mirkok> so V8 is really damn fast
12:23 < jesusabdullah> interesting for sure
12:39 < dominictarr> jesusabdullah: hey whats up?
12:39 < dominictarr> hows india?
12:54 < substack> powered through some testling-ci bugs, now time for voxel-engine fun
12:56 < substack> chrisdickinson: oh sweet, a testling-ci badge for aabb-3d!
12:59 < substack> dominictarr: plane tickets are all set!
12:59 < dominictarr> sweet!
12:59 < substack> looks like I need to buy some $20 e-visa thing though
12:59 < jesusabdullah> dominictarr: not bad!
13:00 < dominictarr> cool, tell marak I say hi
13:02 < dominictarr> jesusabdullah are you gonna merge this https://github.com/jesusabdullah/node-ecstatic/pull/58
13:03 < dominictarr> hij1nx: ping?
13:09 < jesusabdullah> dominictarr: uhh
13:09 < jesusabdullah> dominictarr: lemme take a look here
13:09 < jesusabdullah> dominictarr: I must've missed this with the crazy internet outages here
13:10 < dominictarr> just adds a command so you can start a static server without writing any code
13:10 < jesusabdullah> dominictarr: I approve of the concept, I was considering writing such a thing myself (probably copy-paste from nodejitsu's http server)
13:10 < dominictarr> and you can npm install -g
13:12 < dominictarr> oh fuck, have you guys seen the new mega upload?
13:12 < dominictarr> I've been hearing a bit about it, but I just discovered what it actually is...
13:12 < jesusabdullah> i have not
13:12 < dominictarr> ENCRYPTED FILE SHARING
13:12 < LOUDBOT> I'M SO ANGRY I OVERFLOWED MY HATE BITS AND I'M HAPPY AGAIN
13:13 < dominictarr> http://mega.co.nz/#/mega
13:13 < jesusabdullah> I'm checking out the ole' http-server to make sure there isn't any killer feature you forgot or something XD
13:13 < jesusabdullah> I approve of the relative simplicity of your PR dominic
13:13 < dominictarr> it encrypts it in the browser, and they never even know what you have uploaded
13:14 < dominictarr> jesusabdullah: thanks
13:15 < jesusabdullah> I didn't know ecstatic still supported the (dir, opts) api XD
13:15 < jesusabdullah> did I backwards compat that? I honestly don't remember
13:15 < jesusabdullah> oh hell yeah I did lol
13:15 < jesusabdullah> nice
13:16 < dominictarr> all the features and options fallthrough via optimist, the only thing I added was the convienience option for dir, as first arg, or current dir
13:18 < jesusabdullah> Yeah
13:18 < jesusabdullah> I'm hoping I chose sane defaults for ecstatic's options
13:18 < jesusabdullah> I'm sure I did ;)
13:18 < jesusabdullah> Am I the only one that promptly forgets his own api's?
13:19 < jesusabdullah> or hers
13:19 < jesusabdullah> I suppose
13:20 < dominictarr> who cares they are already documented
13:20 < dominictarr> jesusabdullah: I forget what features I've implemented all the time!
13:21 < jesusabdullah> that's half of why I document it tbh
13:21 < dominictarr> AGREE
13:22 < substack> I would have to be one of those savants to remember all my apis
13:22 < jesusabdullah> dominictarr: published
13:22 < substack> so many
13:22 < dominictarr> sweet!
13:22 < substack> dominictarr has nearly as many modules as I do so it's probably similar
13:22 < jesusabdullah> Yeah, if it's urgent you have to flag me down lol
13:22 < jesusabdullah> I guess  ecstatic surprises me just because it's one of my more polished modules
13:23 < jesusabdullah> you'd think having spent that much time with it I'd remember a thing or two
13:23 < substack> jesusabdullah: I use ecstatic so much
13:23 < substack> pretty much all the time
13:23 < jesusabdullah> Hah!
13:23 < dominictarr> I want a fast way to npm depreciate all the modules I'm I know I'm never gonna use/touch again
13:23 < substack> I don't even know what the api is aside from ecstatic(dir)
13:23 < substack> since that's the only api endpoint I ever use
13:23 < jesusabdullah> It *is* like the only reasonably okay static middleware module that isn't bundled with a bunch of other stuff
13:24 < jesusabdullah> That's funny substack because I'm like 80% sure I "deprecated" that api, as in, it's undocumented
13:24 < dominictarr> jesusabdullah: does everyone ride motoscooters in goa?
13:24 < jesusabdullah> I was all like, "here I'll make it backwards compat because it's easy and it won't break everything"
13:24 < jesusabdullah> dominictarr: scooters, and lots of 150cc-250cc motorcycles
13:25 < jesusabdullah> "motorbikes" I guess
13:25 < dominictarr> cool
13:25 < jesusabdullah> Yeah
13:25 < jesusabdullah> I'm rollin' around on a 150cc hondog
13:25 < dominictarr> the step-though kind?
13:25 < jesusabdullah> a "hero honda"
13:25 < jesusabdullah> no, like a "real" motorcycle
13:25 < jesusabdullah> there are the step-throughs though, that's what I meant by "scooter"
13:25 < dominictarr> there is two levels
13:26 < dominictarr> there is like, scooter type, where you can touch your ankels
13:27 < dominictarr> then there is another type that doesn't go down all the way, but you can stiff step through
13:27 < dominictarr> with enough room for a bag of stuff, or a small child, etc to stand there
13:27 < dominictarr> this is the type that seems most popular here in veitnam
13:28 < dominictarr> but you have a regular "man style" motorbike?
13:28 < jesusabdullah> dominictarr: http://www.tuffcitypowersports.com/assets/uploads/images/scooters/Saga-Quest-Scooter-white.JPG like this but shittier
13:29 < jesusabdullah> and yeah it's a "regular man style motorbike" whatever that means XD
13:30 < dominictarr> right
13:30 < dominictarr> I mean petrol tank is between your thighs
13:32 < jesusabdullah> I think I might have put process.logging support in ecstatic??
13:49 < jesusabdullah> oh good, that was a different project
13:50 < jesusabdullah> I've spent an obnoxious amount of time hooking into the request module's logging mechanism ;)
14:10 < dominictarr> http://dominictarr.github.com/r-edit/
14:10 < pikpik> Goood moooooorning!
14:12 < dominictarr> cianomaidin_: http://dominictarr.github.com/r-edit/
15:08 < dominictarr> todo: create a socket.io2stream module
15:09 < dominictarr> to present the socket.io api, with events and callback craph
15:09 < dominictarr> but streamable
15:09 < dominictarr> to make it easy to port socket.io apps to streams
15:23 < dominictarr> in particular, I want this for tty.js
15:35 < jesusabdullah> in a sense the opposite of your stream interface for socket.io
15:35 < jesusabdullah> or the counterpart
15:35 < jesusabdullah> yeah?
15:40 < pkrumins> yo
15:40 < dominictarr> so like you would go io.connect()
15:40 < dominictarr> and then sock.pipe(io.createStream()).pipe(sock)
15:41 < dominictarr> so it's socket.io without the socket.io
15:41 < dominictarr> so it works over any stream
15:41 < dominictarr> hmm, this is a great troll module, I think I'll do this tomorrow!
15:57 < jesusabdullah> a'ight I'm peacin' out for the night I think
15:57 < jesusabdullah> gnight erryone
15:57 < jesusabdullah> or good day as it may be
16:07 < defunctzombie> substack: with tap-parser are you capturing the test name information?
16:09 < substack> yes
16:18 < defunctzombie> substack: didn't see that in the sample js output in the readme, (thus I wondered)
16:18 < defunctzombie> also, damn you are up early
16:30 < defunctzombie> substack: seen this? https://github.com/eriwen/javascript-stacktrace
16:53 < defunctzombie> isaacs: found another issue with npm (me thinks)
16:53 < isaacs> defunctzombie: oh?
16:53 < defunctzombie> when using the shortform github style urls in dependencies, npm seems to check the registry and fails if it cannot find the module
16:54 < defunctzombie> and doesn't continue and try the install from github
16:54 < defunctzombie> if I use the longform url style it works no problem
16:54 < defunctzombie> this is a problem since my module doesn't exist in the registry yet but I still want to install from github (ideally with shortform url)
16:56 < isaacs> defunctzombie: seems ot work fine on 1.2.2
16:57 < isaacs> what version of npm are you using?
16:57 < defunctzombie> let me check
16:57 < defunctzombie> 1.1.70
16:57 < defunctzombie> with node 0.8.16
16:57 < defunctzombie> let me bump to node 0.8.17 and see if it still fails
16:58 < isaacs> defunctzombie: npm i -g npm
16:58 < defunctzombie> yea, but if it doesn't work with a shipped node version I am less inclined to use it :)
16:59 < defunctzombie> and will just use the longer form
17:07 < isaacs> defunctzombie: it'll work with 0.8.18
17:07 < defunctzombie> k
17:07 < isaacs> defunctzombie: npm updates are not quite 100% in sync with node updates.
17:07 < isaacs> they're a bit more frequent
17:07 < defunctzombie> yea, however most people will just use the bundled npm
17:07 < defunctzombie> unless they *must* have some feature
17:09 < tanepiper> ZOMG http://www.amazon.co.uk/Ready-Player-One-Ernest-Cline/dp/0099560437
17:09 < tanepiper> Like most of humanity, Wade Watts escapes this depressing reality by spending his waking hours jacked into the OASIS, a sprawling virtual utopia where you can be anything you want to be, where you can live and play and fall in love on any of ten thousand planets. And like most of humanity, Wade is obsessed by the ultimate lottery ticket that lies concealed within this alternate reality: OASIS founder *James Hallid
17:09 < tanepiper> ay*, who dies with no heir, has promised that control of the OASIS - and his massive fortune - will go to the person who can solve the riddles he has left scattered throughout his creation.
17:09 < tanepiper> AND SO IT WAS FORETOLD
17:09 < LOUDBOT> AT THE END OF IT ALL, YOU'RE STILL MY BEST FRIEND
17:14 < isaacs> tanepiper: substack WWOOOOAAWWWW
17:15 < tanepiper> isaacs: apparently another character in the book.....
17:15 < tanepiper> Ogden Morrow
17:15 < isaacs> nice!
17:15 < tanepiper> *MIND BLOWN*
17:15 < LOUDBOT> JESUS CHRIST IF THIS QA GUY WAS AT GOOD AT COMING UP WITH TESTS AS HE IS AT SUGGESTING BUG FIXES, WE'D NEVER SHIP A BUG
17:15 < isaacs> OMG IS OASIS A VOXEL UNIVERSE!?!!?
17:15 < LOUDBOT> HELLO I WILL DANCE FOR MONEY$$$$$$
17:16 < isaacs> if so, i'm so gonna rock that shit because it'll be in JavaScript and i'm a JavaScripter
17:20 < ralphtheninja> isaacs: +1 for npm i -g npm, perfect dogfooding :)
17:31 < tanepiper> Yo Dawg....
17:44 < rclosner> mbalho: ping
17:47 < rclosner> mbalho: seems like it would make sense to move block creation/deletion logic into game.js. were you planning on allowing other actions on 'mousedown'?
17:50 < rclosner> mbalho: block create/delete seems universal enough to make it a default (as opposed to having to specify it in each game file).
17:53 < rclosner> mbalho: also, the example in the README won't render by itself. there's an error in Game#showChunk. chunk.position is undefined?
17:54 < rclosner> mbalho: (the voxel-engine README)...
18:09 < isaacs> ralphtheninja: npm's first mission in life has always been to install npm
18:09 < rclosner> mbalho: also, would it make sense to extract voxel textures into its own module? it would be nice if textures came pre-packaged with the engine.
18:21 < mbalho> rclosner: yea weve talked about texture modules
18:22 < mbalho> rclosner: not a bad suggestion about the events stuff, i think some defaults would be nice but its important to keep everything able to be overwritten
18:27 < rclosner> mbalho: agreed. ease of installation is so-so atm. block create/delete should be a default. pointerLock should be called in the engine. also, i'll take a look at the default generator function in the engine. it doesn't work out the box yet.
18:28 < mbalho> rclosner: there isnt a default generator actually
18:28 < rclosner> mbalho: i mean game#showChunk.
18:28 < mbalho> rclosner: ohh right
18:28 < pkrumins> defunctzombie_zz: your range patch landed today in ci testling
18:29 < shama> rclosner: I imagine there will be a good deal of games that dont want to break blocks. For instance a fighter drone game... it wouldnt need clicking to break blocks.
18:29 < rclosner> mbalho: what do you think about using a default generator? setting a convention (in this case) seems worthwhile
18:30 < mbalho> rclosner: makes sense, a little 1 liner for the sphere world would make sense
18:30 < rclosner> shama: that's fair. i wouldn't be surprised if the predominant number of games are going to use block create/delete (given that it's a minecraft derivative). so long as it's overwritable, i don't see a problem with setting it as a default.
18:30 < mbalho> https://github.com/maxogden/voxel/blob/master/index.js#L37
18:31 < mbalho> rclosner: i also wanna make games where clicking edits multiple blocks based on scripting what happens
18:31 < mbalho> rclosner: but that will just override the default
18:32 < rclosner> mbalho: could we pull block creation/deletion into a module? if it's required it will override the 'mousedown' event, if not...
18:33 < mbalho> rclosner: ooh interesting
18:33 < rclosner> mbalho: it's not that much code, but it would create a transparent override methodology.
18:33 < mbalho> rclosner: yea im thinking a module for interacting with blocks in general would be useful as a pattern
18:33 < defunctzombie> pkrumins: \o/ fuck yea
18:34 < defunctzombie> now for https badges and I will be all set for awesome :)
18:34 < rclosner> mbalho: how are you planning on handling textures?
18:35 < mbalho> rclosner: https://github.com/shama/voxel-texture and modules for texture packs that install the pngs
18:35 < rclosner> shama: awesome!
18:35 < shama> mbalho: just sent a PR for that :) https://github.com/maxogden/voxel-engine/pull/4
18:36 < mbalho> woooot
18:36 < mbalho> shama: was implementing that difficult?
18:36 < mbalho> shama: me and substack were stumped
18:37 < rclosner> pointerLock make sense as a default? trying to minimize the amount of code one needs to get off the ground.
18:37 < pkrumins> defunctzombie: i'll get https badges right after i launch ie10s for browserling
18:37 < defunctzombie> ooooo fancy
18:37 < mbalho> rclosner: yea for now its required
18:38 < mbalho> rclosner: we could do a mouse drag control for people without pointerlock but im gonna wait until we port all this to mobile to implement that
18:38 < shama> mbalho: it wasnt actually... I was pretty skeptical about it. Turns out each face has a Vector3 normal. The normal points in the direction the face is facing. So just a matter of selecting the correct materialIndex based on the normal.
18:38 < mbalho> shama: wow nice
18:38 < shama> Here is the line: https://github.com/shama/voxel-texture/commit/cd507cee624e2dd273b6608b435b5a32eac59d39#L2R64
18:39 < rclosner> mbalho: it would be awesome to get down to something like this: https://gist.github.com/4567043
18:39 < mbalho> rclosner: we could default to document.body :)
18:39 < mbalho> shama: that is awesome
18:41 < shama> still dont know why back and left are rotated though... at the very least I can use canvas to rotate those textures upon loading but I want to find out why they do that.
18:41 < mbalho> woot go to 30 seconds in http://www.ustream.tv/recorded/28597081
18:41 < mbalho> (voxel.js demo)
18:42 < shama> nice! :)
18:44 < rclosner> mbalho: word.
18:44 < CoverSlide> fuck yeah
18:45 < defunctzombie> pkrumins: you guys should hook up with sendgrid to give email notifications on failures
18:45 < CoverSlide> \o/
18:48 < pkrumins> awesome idea
18:48 < rclosner> mbalho: also, API weirdness? game.voxels.chunks (reverse hierarchy).
18:49 < defunctzombie> mbalho: have you run node on r-pi yet?
18:50 < defunctzombie> pkrumins: https://npmjs.org/package/emailjs I have used that for quite some time now to send text and html emails with great success
18:51 < defunctzombie> I have basic html email templates if you guys are interested (found them online at some point) and they seem to do well in most viewers
18:55 < pkrumins> cool, i'll add email sending!
18:56 < mbalho> defunctzombie: yep
18:57 < mbalho> rclosner: yea we put the chunking logic into the voxel module so that all that crap is in one place
18:57 < mbalho> rclosner: so game.voxels = require('voxels') and so you have to do game.voxels.chunks
18:57 < CoverSlide> oh yeah? how's the frame rate?
18:57 < mbalho> rclosner: there is probably a better way to represent that stuff
18:57 < mbalho> defunctzombie: i reallyr eally really really wanna get firefox os running on the rpi
18:57 < mbalho> defunctzombie: it was done back in august
18:57 < mbalho> defunctzombie: then i can run the voxel game on it!
19:00 < rclosner> mbalho: ah, duh. i figured out the bug. README example doesn't work because voxel and game.voxel are different. no chunk positions defined in voxel.
19:01 < rclosner> mbalho: works if you do something like this: https://gist.github.com/4567281
19:08 < defunctzombie> mbalho: nice, I want to run some raw http server nodejs benchmarks on an r-pi
19:08 < defunctzombie> mbalho: to see what type of cpu/network perf one can get out of it
19:08 < defunctzombie> for serving basic web apps
19:13 < rclosner> mbalho: thoughts on extracting functions like scale() into a voxel-utils repo?
19:19 < ehd> defunctzombie: are you on raspbian?
19:20 < defunctzombie> ehd: I am on nothing yet, just got the r-pi from a friend, but he said to give raspbian a go
19:20 < defunctzombie> I will most likely do that since I run debian normally
19:21 < defunctzombie> I need it headless tho so need something with ssh out of the box
19:21 < CoverSlide> i suggest checking out arch
19:21 < defunctzombie> is that an rpm offshoot?
19:21 < CoverSlide> no, it's got its own package manager
19:21 < defunctzombie> I won't be switching away from debian systems anytime soon :)
19:21 < CoverSlide> so no deb
19:21 < CoverSlide> but the arch build system is awesome
19:22 < defunctzombie> little gain for too much relearning :/
19:22 < defunctzombie> is it like gentoo where I need to build everything? if so, no thanks :D
19:22 < CoverSlide> the gain is actually enormous
19:22 < CoverSlide> no, it has binary packages
19:22 < defunctzombie> debs such to make, but thankfully I don't have to make them anymore
19:23 < defunctzombie> I just run debian since that is what I run on my servers
19:23 < CoverSlide> but it's fairly easy to build your own if you want an updated version
19:23 < defunctzombie> maybe one day I will explore other things but for now it hasn't made it to the top of the list
19:23 < CoverSlide> alrighty
19:23 < defunctzombie> I kinda go with a good enough approach at the OS level these days :/
19:24 < defunctzombie> and only bring in updates of things I really need or want
19:26 < CoverSlide> well if you have time, i recommend reading up on arch. it's different from deb, but i made the switch and am never looking back.
19:29 < Raynos> defunctzombie: make a PR and it should support the entire lookup algorithm
19:29 < CoverSlide> there is an issue with stability, which once in a blue moon there is a breaking change, which they will usually announce on the homepage, so for servers, it's definitely not a great choice, but for my laptop and toy boxes it's awesome
19:30 < mbalho> rclosner: what else would be in a voxel-utils module
19:30 < mbalho> rclosner: we could also add stuff to the voxel module for now and spin out another module later if it makes sense
19:32 < rclosner> mbalho: sounds good to me. since you were talking about extracting lat/lng conversion logic into a separate module, i was thinking utils might be a good place for that.
19:33 < jaekwon1> lat/lng… could be its own module for geo logic, including maybe geo-quading in the future
19:38 < rclosner> jaekwon1: geoquading? is that like creating a bounding box from a point of origin?
19:39 < jaekwon1> rclosner: it's tiling up the lat/lng of earth for simple indexing
19:40 < jaekwon1> then when you store some object or place, you can index both the lat/lng of the place as well as the geoquad identifier
19:41 < rclosner> jaekwon1: got it. thanks.
19:44 < mbalho> heres a module from 3 years ago that i made to do stuff with geojson https://github.com/maxogden/geojson-js-utils
19:46 < mbalho> rclosner: something like o/ would be useful i suppose
19:54 < Raynos> defunctzombie: https://github.com/Raynos/tools/issues/1
20:14 < mbalho> just did a venturebeat interview about voxel.js woowoo they said they wanted to feature the project
20:15 < CoverSlide> awersome
20:21 < jaekwon1> mbalho: what is checkBlock? why does it p.x -= 1.1 * d.x ?
20:21 < mbalho> jaekwon1: substack wrote that but it checks if you can place a block
20:22 < mbalho> jaekwon1: i think it looks adjacent to the block that you clicked on to see if you can place a block there and says no if the player is in the way etc
20:22 < jaekwon1> oh ok
20:29 < mbalho> copy paste this bad boy into a voxel demo https://gist.github.com/4499643
20:30 < mbalho> chrisdickinson: whoa your collision demo is awesome
20:30 < chrisdickinson> thanks!
20:31 < chrisdickinson> i'm gonna spend some more time on it this weekend and iron out the rest of the kinks
20:31 < mbalho> chrisdickinson: thats awesome. looking forward to the pull req
20:32 < mbalho> chrisdickinson: also gravity in your demo so funny ahahaha
20:32 < chrisdickinson> yeah, it's weird! jumps are instantaneous, but gravity isn't
20:32 < chrisdickinson> http://github.com/chrisdickinson/collide-3d-tilemap is the api i'm using to do the collisions
20:33 < chrisdickinson> i need to update the readme though, it's kind of a clone of collide-2d-tilemap
20:33 < chrisdickinson> https://github.com/chrisdickinson/collide-3d-tilemap/blob/master/index.js for the relevant code
20:35 < isaacs> defunctzombie_zz: 0.8.18 coming soon
20:38 < defunctzombie> isaacs: \o/
20:38 < isaacs> compiling on the proprietary operating systems, still
20:38 < isaacs> they're the slowest
20:39 < isaacs> darwindows
20:44 < defunctzombie> haha
20:44 < defunctzombie> darwindows amazing
20:45 < defunctzombie> Raynos: am I meant to do something with this repo?
20:46 < CoverSlide> Linux++
20:48 < CoverSlide> why use something fast and simple like GCC when you can have XCode or VS take up 4G on your disk?
20:48 < defunctzombie> Raynos: looking at the issue for npm-css I would say that prefix-css, while interesting, is out of scope since the dev still has to prefix their stuff in their templates. Seems better to just be explicit in the css. Yes, it is a bit repetitive but :/
20:57 < defunctzombie> CoverSlide: why use something old and slow like gcc when you can have clang :p
21:04 < mbalho> WOOT https://twitter.com/voxeljs/status/292376278064037888
21:06 < guybrush_> mbalho: which device?
21:06 < mbalho> guybrush_: galaxy nexus
21:07 < guybrush_> sweeeet
21:07 < guybrush_> i need to get such phones from somewhere
21:07 < mbalho> im gonna test on the nexus 7 as soon as mine charges
21:08 < guybrush_> basically i would need one without any contract
21:08 < CoverSlide> where you get bleding edge chrome?
21:08 < guybrush_> playstore :D
21:08 < CoverSlide> i just have beta chrome
21:08 < guybrush_> thats all you need
21:08 < CoverSlide> oic
21:08 < guybrush_> https://twitter.com/alteredq/status/292069924128563200
21:09 < jaekwon1> mbalho: player-phsyics is player coordinate centric. it should be world centric.
21:09 < CoverSlide> ohh nice
21:09 < jaekwon1> mbalho: freedom constraints should be constraints should be constraints on world coordinates
21:09 < guybrush_> basically nowadays we live in the future!
21:10 < mbalho> jaekwon1: chrisdickinson is working on a new version
21:10 < guybrush_> its like everyday more is possible
21:10 < jaekwon1> oh cool. is there a demo?
21:10 < mbalho> jaekwon1:  yea neversaw.us/voxel/
21:11 < CoverSlide> jumping is weird
21:11 < jaekwon1> jumping doesn't jump, it teleports back in time!
21:12 < mbalho> hehe
21:12 < mbalho> its a demo
21:12 < mbalho> not finished :)
21:12 < jaekwon1> ok sweet, it looks good
21:15 < defunctzombie> pkrumins: another feature idea, once you know what browsers the test will run under, make pending placeholders on the webpage
21:15 < defunctzombie> pkrumins: otherwise it isn't so clear to me if you are gonna run tests or if it is done
21:17 < rclosner> chrisdickinson: looks awesome!
21:20 < pkrumins> defunctzombie: noted
21:21 < defunctzombie> pkrumins: http://ci.testling.com/shtylman/dom <-- using ranges now :)
21:21 < defunctzombie> really awesome
21:21 < defunctzombie> package.json so much simpler
21:21 < defunctzombie> https://github.com/shtylman/dom/commit/8cf7b0e1d53ffe2444f1bdfba4de45ab65080543
21:24 < defunctzombie> substack: pkrumins: http://ci.testling.com/shtylman/dom.png seems there is an ordering issue on the badge
21:33 < substack> ranges aren't all the way fixed I think
21:35 < defunctzombie> probably not, also, my tests seem to have hung :(
21:38 < pkrumins> it didnt run any tests past chrome 8
21:40 < pkrumins> and didnt attempt to run safari or opera tests
21:40 < defunctzombie> sadness
21:41 < defunctzombie> was it a failure with ranges or with something else?
21:41 < pkrumins> it didnt crash
21:42 < pkrumins> it just stopped after chrome 8
21:42 < pkrumins> i dont know why it stopped
21:43 < defunctzombie> the firefox range worked all the way, same with ie
21:43 < pkrumins> yeah
21:44 < pkrumins> can you run the same test again?
21:44 < pkrumins> just trigger the hoook
21:44 < pkrumins> i awnt to see if it gets past chrome 8
21:46 < defunctzombie> k
21:46 < pkrumins> seeing the commit
21:46 < pkrumins> following it
21:50 < Raynos> defunctzombie: prefix-css is about being able to import someone else's css under a namespace
21:50 < Raynos> which means you don't have to worry about them polluting your global css namespace because they may be naughty
21:50 < Raynos> defunctzombie: also that was my original idea about npm-css so talk about npm css related stuff there or on the repo
21:51 < Raynos> isaacs: Am I supposed to rebase the objects stuff yet?
21:51 < fotoverite> How goes the jitsu invasion this week mbalho?
21:51 < isaacs> Raynos: GO FOR IT!  GO FOR THE GOLD!
21:51 < defunctzombie> Raynos: gotcha, yea, I could see that aspect
21:51 < isaacs> Raynos: it would be good to keep thatpull req focused on object streams, though.  the other stuff makes it harder to evaluate.
21:52 < isaacs> Raynos: if there are other patches that need to land to make object streams work, then fine.
21:52 < isaacs> but if it's just generic stream bugs, it should be separate.
21:52 < Raynos> isaacs: i need to land those patches to get the tests to run. Didn't want a PR that didn't pass tests
21:52 < isaacs> sure
21:52 < defunctzombie> Raynos: maybe if there was a good way to test it and have the css stuff applied automatically during testing (manual) or such then it would be better
21:53 < Raynos> anyway
21:53 < isaacs> Raynos: also, don't do that silly tap-alike thing.  i only did that to be able to more easily copyu over the tests i already had in readable-stream
21:53 < isaacs> Raynos: but now i regret it
21:53 < Raynos> I will rebase it this weekend
21:53 < defunctzombie> Raynos: I would like to explore the idea of required prefixes cause if you do that from the start
21:53 < defunctzombie> then no one can make mistakes
21:53 < isaacs> Raynos: awesome :)
21:53 < defunctzombie> by forgetting
21:53 < Raynos> isaacs: I like the tap-alike thing :p
21:53 < isaacs> yeah.
21:53 < isaacs> but it's overly cute.
21:53 < Raynos> i dont know how to async test without t.end()
21:53 < isaacs> node tests should just throw or not throw
21:54 < isaacs> Raynos: you just do a thing, and then it either throws or it doesn't.
21:54 < Raynos> then each test needs a seperate file
21:54 < isaacs> Raynos: and you have a process.on('exit',fn) that makes sure the test actually ran
21:54 < isaacs> yeah
21:54 < isaacs> well, you can do a bunch of tests in one file, still
21:54 < isaacs> we do that a lot
21:54 < Raynos> I dont know how to do MULTIPLE t.end()'s in one file
21:54 < isaacs> look at the other tests in test/simple/
21:54 < isaacs> almost all of them do multiple things
21:54 < Raynos> well its basically a counter thing
21:54 < Raynos> which is what the tap-alike does
21:55 < defunctzombie> how does that work when writing browser oriented tests? (whatever you guys are talking about)
21:55 < isaacs> foo(function() { fooRan = true }); process.on('exit', function() { assert(fooRan) }); bar(...)
21:55 < isaacs> etc.
21:55 < Raynos> defunctzombie: it doesn't
21:55 < isaacs> defunctzombie: we're not talking about browser-oriented tests at all :)
21:55 < defunctzombie> ah :)
21:55 < Raynos> defunctzombie: with browsers you need counters and tape
21:56 < pkrumins> defunctzombie: it looks like it stopped again at chrome 8.
21:56 < defunctzombie> pkrumins: fascinating
21:58 < defunctzombie> pkrumins: nope, seems to be going
21:58 < pkrumins> ah no it just run chrome 9
21:59 < defunctzombie> Raynos: actually, the more I think about it, the more I think auto prefix should be a requirement...
21:59 < defunctzombie> Raynos: just have to sell it right
21:59 < defunctzombie> pkrumins: maybe it was some one time issue?
22:00 < pkrumins> not sure
22:01 < Raynos> defunctzombie: you need some kind of namespacing.
22:01 < Raynos> defunctzombie: just like javascript exports a single function. CSS should only use a single global name
22:01 < defunctzombie> Raynos: yea, I would agree
22:01 < Raynos> then you can build a tool to verify no name conflicts
22:02 < Raynos> BUT
22:02 < defunctzombie> the idea of prefixing by name of module sounds reasonable
22:02 < Raynos> you want to be able to do `require "../lib/bootstrap.css" as PREFIX`
22:02 < defunctzombie> hm, would you?
22:03 < defunctzombie> I think that is a different case than the widget namespace issue
22:04 < defunctzombie> as a widget author you should be able to know what namespace will be used
22:04 < Raynos> well no
22:04 < Raynos> i dont want boostrap
22:04 < Raynos> bootstrap is a diaster
22:04 < Raynos> defunctzombie: if you use a different namespace for a widget
22:04 < defunctzombie> Raynos: yea, then you would not include it right?
22:04 < Raynos> then its your responsibility to do `addClass(widget.view, "namespace")`
22:05 < Raynos> but I think in the transitional thing
22:05 < defunctzombie> meh, I dunno seems like it is started to get more complex and involved
22:05 < defunctzombie> you may not be able to just addClass(widget.view) or whatever
22:05 < defunctzombie> user shouldn't have to do that :/
22:05 < Raynos> the transitional period
22:05 < pkrumins> looks like it stopped at chrome 22
22:06 < Raynos> I may want to prefix other peoples css
22:06 < Raynos> I guess I can just import their css and prefix it myself
22:06 < Raynos> meh this problem shouldnt be solved in this tool
22:06 < Raynos> just require is fine
22:07 < defunctzombie> I don't think general namespacing problems should be solved
22:07 < defunctzombie> the only namespace thing I think would be reasonable is the idea that all your stuff must be namespaced
22:08 < defunctzombie> and doing it for you with teh module name seems clear and consistent
22:08 < defunctzombie> in the widget I was playing around with I did it all manually
22:09 < defunctzombie> but imagine if someone decides to add "p { }"
22:09 < defunctzombie> or something
22:09 < defunctzombie> that is really annoying certainly, now we can educate and inform that you shouldn't do these things
22:09 < defunctzombie> but maybe auto namespacing is a nicer approach
22:10 < defunctzombie> pkrumins: again with no error?
22:10 < defunctzombie> pkrumins: chrome 22 worked before btw
22:10 < pkrumins> yes
22:10 < defunctzombie> so I am pretty sure it isn't an issue with failing tests
22:11 < pkrumins> it's not an issue with failing tests, it's something else
22:11 < pkrumins> they just stop being run
22:11 < defunctzombie> haha
22:11 < defunctzombie> maybe they get tired
22:11 < pkrumins> i've seen it before
22:11 < pkrumins> when lots of browsers are used then it happens
22:11 < pkrumins> probably they got tired yes.
22:11 < pkrumins> i think we forgot to require('redbull')
22:16 < defunctzombie> haha
22:19 < defunctzombie> Raynos: basically, if you know your stuff will be namespaced by module name then you make templates and such accordingly
22:19 < defunctzombie> also, simulated shadow dom a bit better I think
22:19 < Raynos> defunctzombie: yes BUT you still need a way to enforce that two modules cant use smae namespace
22:19 < Raynos> but that's a seperate tool!
22:19 < Raynos> maybe it belongs in npm-css though
22:19 < defunctzombie> Raynos: seems less likely to happen.. you would have to install both of them with the same name
22:20 < Raynos> oh wait!
22:20 < defunctzombie> they don't get a choice in namespace
22:20 < Raynos> AUTO NAMESPACE WITH PACKAGE NAMe
22:20 < defunctzombie> yes
22:20 < Raynos> NPM ENFORCES ZERO CONFLICTS
22:20 < LOUDBOT> DAMN IT WHY DO ALL THE HOT WOMEN HAVE TO BE OVER 40?
22:20 < Raynos> FUCK YES
22:20 < LOUDBOT> GUYS I THINK YOU SHOULD KNOW THAT I HAVE HAD 6 OF SANTA'S REINDEER LOCKED UP IN MY BASEMENT FOR THREE MONTHS NOW SO... YOU KNOW.  SHUT IT DOWN.
22:20 < defunctzombie> that is what I have been saying :)
22:20 < Raynos> GLOBAL NAMESPACE FUCK YES.
22:20 < LOUDBOT> JOHNMMMMMMM
22:20 < Raynos> defunctzombie: DO IT. LAND IT
22:20 < Raynos> isaacs: the fact that npm has a global namespace is best thing ever.
22:20 < defunctzombie> hahaha
22:21 < defunctzombie> substack: is css-prefix still alive?
22:21 < defunctzombie> or is there an alternate lib?
22:21 < Raynos> defunctzombie: theres an alternative
22:22 < Raynos> defunctzombie: https://github.com/visionmedia/rework#prefixselectorsstring
22:24 < defunctzombie> Raynos: any reason to use that over just css-prefix which does that one thing?
22:25 < defunctzombie> rework is cool in the general case
22:25 < Raynos> css-prefix prefixes EVERYTHING
22:25 < Raynos> which is bullshit
22:25 < Raynos> you just want to prefix the selectors with another class selector
22:25 < Raynos> prefixing all your classes in your template sucks
22:25 < defunctzombie> Raynos: you have to prefix everything
22:25 < Raynos> all you want to do is add one class to your root
22:25 < Raynos> no
22:25 < Raynos> you add one class to your root
22:26 < defunctzombie> ?
22:26 < Raynos> <div class="my-module">{{ANYTHING}}</div>
22:26 < Raynos> and you any css selector for anything in {{ANYTHING}} will work
22:26 < Raynos> because of prefixselectorstring
22:26 < defunctzombie> and for that to be correct you need to prefix everything
22:26 < defunctzombie> otherwise if I define "p { ... }" in my css
22:26 < defunctzombie> I will mess up your page
22:27 < Raynos> no it wont
22:27 < Raynos> because if you prefix it
22:27 < Raynos> then its .my-module p { ... }
22:27 < defunctzombie> yes.. which means you need to prefix everything
22:27 < defunctzombie> with .my-module
22:27 < Raynos> yes agreed
22:27 < Raynos> thats not what css-prefix does
22:27 < Raynos> it prefixes every class and id with a string
22:28 < defunctzombie> oh.. yea
22:28 < Raynos> and doesn't prefix every selector with a string
22:28 < defunctzombie> shit
22:28 < defunctzombie> that is wrong
22:28 < defunctzombie> I thought it did the other way
22:28 < defunctzombie> teh way we needed :)
22:28 < Raynos> just steal the prefixSelector out of rework
22:28 < Raynos> make it a module
22:28 < Raynos> and PR rework
22:28 < defunctzombie> haha
22:28 < Raynos> and leave some kind of comment to TJ
22:28 < Raynos> telling him he is doing it wrong
22:29 < defunctzombie> maybe substack is not using css-prefix and wouldn't mind the change I dunno
22:29 < defunctzombie> he is probably using it
22:35 < chrisdickinson> rclosner: belatedly, thanks!
22:35 < chrisdickinson> jaekwon1: yep, the player-physics being in model space is a pain. rewrite will put it in world space, ideally
22:36 < Raynos> dominictarr: https://github.com/dominictarr/crdt/pull/18
22:36 < Raynos> defunctzombie: I think the parentClass property might work as we expect
22:37 < defunctzombie> Raynos: of rework? or something else?
22:38 < mbalho> chrisdickinson: i think playerphysics shouldnt translate the yawObject
22:38  * chrisdickinson nods
22:38 < mbalho> chrisdickinson: that should be done by whatever is using player physics
22:38 < chrisdickinson> i haven't actually looked at how the yawObject is used in rendering
22:38 < mbalho> chrisdickinson: i implemented super simple collisions that arent as good as yours but they are very straightforward https://github.com/maxogden/voxel-engine/commit/72fed96c23167d8a0ab0b2318c4015d67336b20e
22:38 < Raynos> defunctzombie: https://github.com/substack/css-prefix#insertprefixopts-src parentClass
22:39 < mbalho> chrisdickinson: its just an experiment, im not gonna use it but it made me realize that the physics module should only set velocity
22:39 < mbalho> chrisdickinson: and whatever consumes the physics module should set position
22:39 < chrisdickinson> mbalho: if player physics can A) present a vector in world space and B) accept a world space vector and transform it back to model space, that would be ideal
22:40 < chrisdickinson> or yeah, manually setting position. that's what my demo does.
22:40 < chrisdickinson> (it actually zeroes out velocity and sets position manually)
22:40 < defunctzombie> Raynos: yep seems like it will
22:40 < defunctzombie> I will have to try it out
22:40 < defunctzombie> I think we are getting closer :)
22:41 < Raynos> FFFFFFFFFFFFFFFFFFFFf
22:41 < mbalho> chrisdickinson: if you clone the collisions branch and then comment out the last 3 lines of index.js in player physics (this.yawObject.translateX( this.velocity.x ) etc) then it should work
22:41 < mbalho> chrisdickinson: after npm installing
22:44 < Raynos> I installed a broken version of npm
22:44 < Raynos> I cant do npm i npm@1.1 :(
22:44 < Raynos> I have to like download shell scripts
22:46 < defunctzombie> haha
22:46 < defunctzombie> that is what you get for trying to mess with npm
22:50 < chrisdickinson> mbalho: these are the modifications to voxel-engine for my demo: https://gist.github.com/da55ec060160d0bf82b5
22:51 < defunctzombie> pkrumins: let me know if you want me to trigger tests at any point
22:53 < chrisdickinson> mbalho: actually, here's it with the patch to player-physics https://gist.github.com/3ec0445a4c8f3aed8084
22:53 < pkrumins> defunctzombie: ok!
22:54 < chrisdickinson> the nice bit is that for object->terrain collisions, we could implement a hook system for "tile-type" -> "function(worldObject)"
22:54 < chrisdickinson> so, if the player hits lava, or water, or something like that
22:54 < chrisdickinson> or if the player hits the ground, play a sound, etc
22:58 < Raynos> defunctzombie: your fork of npm-css seems ok. it just needs more tests
22:59 < defunctzombie> Raynos: agreed
22:59 < defunctzombie> I will add them after the css-prefix stuff
22:59 < defunctzombie> if you can think of specific ones make issues
22:59 < Raynos> also an async version would be nice
22:59 < defunctzombie> on the original project if you want
22:59 < Raynos> but for my use case I actually want a sync version :P
22:59 < defunctzombie> Raynos: I agree, however too many parts of this are sync
22:59 < defunctzombie> like resolve
22:59 < defunctzombie> and such
22:59 < defunctzombie> eventually we will do async
22:59 < defunctzombie> I thought about it
22:59 < Raynos> well resolve probably has an async version
23:00 < Raynos> but `append(document.head, Style(npmcss(require("./foo.css"))))`
23:00 < Raynos> is what I want
23:00 < Raynos> and woh LISP
23:00 < defunctzombie> haha
23:01 < Raynos> substack: https://github.com/substack/tape/pull/14 :(
23:04 < defunctzombie> Raynos: I basically got what I wanted, once I wrote the makeup module haha, but now with prefixes it will be even better
23:05 < Raynos> defunctzombie: I can't transfer the repo when you have a fork :(
23:05 < defunctzombie> :(
23:05 < defunctzombie> want me to delete the fork?
23:06 < defunctzombie> and then transfer
23:09 < Raynos> that works
23:10 < defunctzombie> Raynos: deleted
23:11 < Raynos> requested
23:12 < defunctzombie> do I need to do something now?
23:13 < mbalho> chrisdickinson: cool im gonna spend a couple of hours on voxel-engine now
23:13 < Raynos> defunctzombie: you should get a thing that says accept request
23:13 < Raynos> or I send it to wrong person
23:14 < Raynos> Oh
23:14 < Raynos> > This repository is awaiting GitHub staff approval to transfer ownership to shtylman.
23:14 < Raynos> lol.
23:15 < defunctzombie> haha
23:15 < defunctzombie> amazing
23:15 < defunctzombie> maybe it will be rejected :D
23:19 < chrisdickinson> mbalho: also, i've found it handy to hack voxel-engine into using browservefy instead of watch
23:19 < mbalho> chrisdickinson: cool
23:20 < mbalho> chrisdickinson: what would be a good npm script for it?
23:20 < chrisdickinson> npm install -g browservefy; browservefy demo.js 8000 -- -d
23:20 < mbalho> chrisdickinson: whats the -d do?
23:20 < chrisdickinson> turns on debug output
23:20 < chrisdickinson> i mean, sourceURL stuff
23:20 < mbalho> ahh yea
23:21 < chrisdickinson> (and you'll have to modify demo/index.html to point at "demo.js", not "/dist/browserify-bundle.js"
23:21 < chrisdickinson> )
23:22 < chrisdickinson> browservefy is basically just python -m SimpleHTTPServer, but with the ability to specify a single path as being a browserify entry point (and then after the "--" it forwards all of the following options to browserify cli)
23:22 < chrisdickinson> and it'd actually be `browservefy demo/demo.js 8000 -- -d`, run from the root of the project :D
23:23 < chrisdickinson> https://github.com/chrisdickinson/browservefy/blob/master/index.js#L23-L31
23:24 < pkrumins> wrote a new article guys - how to write testling-ci tests - http://www.catonmat.net/blog/howto-write-testling-ci-tests/
23:25 < mbalho> chrisdickinson: ahh i see so if you do browservefy foo/bar.js 800 you would request localhost:8000/bar.js ?
23:25 < pkrumins> next i'm deploying IE10s for browserling, then getting an ssl cert for testling-ci, then getting Mac OSX for testling and browserling!
23:25 < defunctzombie> \o/
23:25 < pkrumins> mac osx through mac mini colocation! http://macminicolo.net/
23:25 < mbalho> chrisdickinson: so instead of file watcher it just browserifies every time, i see
23:25 < defunctzombie> pkrumins: what is currently running through IE10?
23:26 < pkrumins> defunctzombie: whatcha mean?
23:26 < pkrumins> only testling has ie10 right now
23:26 < defunctzombie> oh.. browseriling.. my bad
23:26 < defunctzombie> :)
23:26 < chrisdickinson> yep!
23:27 < chrisdickinson> basically any request that would hit "foo/bar.js" really calls "browserify <args> foo/bar.js" in a child_process.spawn and pipes it to the response
23:27 < mbalho> chrisdickinson: i dig it, ill switch the npm run watch command to spawn browservefy
23:27 < mbalho> chrisdickinson: + put it in a devdependency
23:27 < mikolalysenko> chrisdickinson: browservefy?
23:27 < chrisdickinson> :D
23:27 < mikolalysenko> https://github.com/mikolalysenko/Serverify
23:28 < defunctzombie> chrisdickinson: why does it do a child process?
23:28 < chrisdickinson> just because it's easy.
23:28 < mikolalysenko> yeah
23:28 < defunctzombie> browserify has an api tho
23:28 < chrisdickinson> and it means that i can forward command line args to browserify
23:28 < chrisdickinson> oh i know
23:28 < mbalho> its doesnt have to be scalabe
23:28 < defunctzombie> ah ok
23:28 < mbalho> scalable*
23:28 < defunctzombie> just making sure you know :)
23:28 < chrisdickinson> this is just for local dev
23:29 < chrisdickinson> might add a switch that automatically spits out a fake index.html if it doesn't see one at the given url, but haven't really decided on that
23:29 < mikolalysenko> chrisdickinson: dang, somehow I missed your project
23:29 < defunctzombie> you totally should :p
23:29 < defunctzombie> chrisdickinson: basically sounds just like zuul
23:30 < defunctzombie> except more generic
23:30 < chrisdickinson> zuul?
23:30 < chrisdickinson> ah!
23:30 < defunctzombie> https://github.com/shtylman/zuul/
23:30 < defunctzombie> everything is run through browserify
23:30 < mbalho> haha everyone has a competing thing
23:30 < defunctzombie> and index.html is served for you since it is mocha specific
23:31 < defunctzombie> not really for "dev" but just to run tests locally before sending them off to testling :)
23:32 < defunctzombie> once this npm-css stuff irons out I will want a server that does npm-css and browserify :) as a test platform haha
23:32 < mbalho> chrisdickinson: the git link in browservefys package.json is wrong
23:32 < chrisdickinson> cripes!
23:33 < mbalho> the value is bad and you should feel bad!
23:33 < chrisdickinson> i do! now!
23:33 < chrisdickinson> will fix
23:34 < mbalho> chrisdickinson: only sucky thing is that its not compatible with workflows like substacks where it goes into bundle.js
23:34 < mbalho> chrisdickinson: but i can just force substack to use it :D
23:35 < chrisdickinson> haha
23:35 < chrisdickinson> yeah, it might be nice to alias
23:35 < chrisdickinson> browservefy path/to/thing.js:bundle.js or something like that
23:36 < mbalho> or just an option
23:39 < mbalho> chrisdickinson: hmm from voxel-engine root dir "browservefy demo/demo.js 8000 -- -d" results in "Not Found" for localhost:8000
23:40 < chrisdickinson> mbalho: remember, your demo is located in the demo/ directory
23:40 < chrisdickinson> so, "localhost:8000/demo/"
23:41 < mbalho> chrisdickinson: ahh. i am used to ecstatic which has a file browser built in
23:41 < chrisdickinson> aaah
23:43 < mbalho> chrisdickinson: can you set the base path?
23:43 < mbalho> chrisdickinson: oh wait nevermind
23:44 < mbalho> chrisdickinson: you have to do localhost:8000/demo/ not localhost:8000/demo
23:44 < mbalho> shama: omg new textures are awesome!
23:44 < mbalho> minecraft eat your heart out!
23:44 < mbalho> haha i should called voxel.js 'yourcraft'
23:45 < mbalho> maybe that will be the name of the voxel.js powered minecraft implementation
23:45 < mbalho> chrisdickinson: i think if browservefy used ecstatic rather than filed it would solve some n00b problems like i just encountered
23:53 < mbalho> chrisdickinson: also i would like to add browserify to devDependencies and be able to give browservefy a path to the browserify bin e.g. ./node_modules/browserify/bin/browserify
23:54 < chrisdickinson> mbalho: could you open those up as issues on the browservefy github?
23:54 < mbalho> sure
23:54 < chrisdickinson> just so i can keep 'em in my head for later when i'm hacking
23:54 < chrisdickinson> thanks!
23:56 < mbalho> substack: whoa i never saw the gather.html demo with the little counters implemented till just now
--- Log closed Sat Jan 19 00:00:49 2013